Author Topic: Unexpected voltage drop on a low current load  (Read 9310 times)

0 Members and 1 Guest are viewing this topic.

Offline fyrewurxTopic starter

  • Contributor
  • Posts: 16
Unexpected voltage drop on a low current load
« on: January 02, 2014, 08:48:39 pm »
Hi everyone,

I am fooling around with a LED strip that is using ucs 5903 controller. I have achieved what I desired (some special Christmas lights). I am posting here because something I just don't get it(probably because I am a newbie): when all the LED's are lit (all white) the voltage drops a lot (at the begining and at the end of the strip). On a cheap dmm it gets to about 2.5 V (from 5V), but I don't have min/max so it can be lower. With the same dmm the max current is 0.7A. Initially I used a 5V/2A power supply, until I used all the LED's (going for white) and the strip went haywire. I thought the supply wasn't clean enough so I added some 3700uF capacitors but the voltage was still dropping. Finally i used a PC power supply (old, about 250W but should give more then 10A on the 5V rail) but the voltage is still dropping (same on dmm but I guess not that low because the ucs 5903 don't mix colors, except the last led). I expect the voltage to be lower at the end of the strip but it's low at the begining too. So, why is the voltage dropping when the power supply should be able to deliver more than 10A on a 0.7A load with more then 1V?

Thank you,

Alex

P.S. Sorry for bad english, and long post.
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #1 on: January 02, 2014, 09:45:41 pm »
For clarity, draw your setup and indicate on the drawing where you are measuring voltages.
~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Offline Jon86

  • Frequent Contributor
  • **
  • Posts: 526
  • Country: gb
Re: Unexpected voltage drop on a low current load
« Reply #2 on: January 02, 2014, 09:49:53 pm »
How long are the cables going to the strip?
Death, taxes and diode losses.
 

Offline fyrewurxTopic starter

  • Contributor
  • Posts: 16
Re: Unexpected voltage drop on a low current load
« Reply #3 on: January 02, 2014, 10:06:57 pm »
Sorry for the crude draw, I hope gets things clear. I know it is not correct but that's not the point of the drawing.

The cable is about 3 meters so I expect a voltage drop at the end of the strip, but why is there a drop across the all circuit (sometimes the BOD of the MCU kicks in and that is close to the supply).
 

Lurch

  • Guest
Re: Unexpected voltage drop on a low current load
« Reply #4 on: January 02, 2014, 10:15:00 pm »
Sorry for the crude draw, I hope gets things clear.

Nope.

I know it is not correct but that's not the point of the drawing.

Yes it is.
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #5 on: January 02, 2014, 10:21:31 pm »
It's not clear if you are saying there are 50 leds, or if there are 50 sets of 3 leds.

Either way, what is the expected current and power.  Show your working, as a maths teacher would say.  Ohm's law is your friend.
~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #6 on: January 02, 2014, 10:24:28 pm »
Also, I suggest actually measuring the resistance of your wire lengths before you go much further.
~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Offline fyrewurxTopic starter

  • Contributor
  • Posts: 16
Re: Unexpected voltage drop on a low current load
« Reply #7 on: January 02, 2014, 10:49:42 pm »
Sorry for the crude draw, I hope gets things clear.

Nope.

I know it is not correct but that's not the point of the drawing.

Yes it is.

Just saying nope, won't help me find what is unclear, is clear for me just hard to put it on text (or drawing). The diagram is not correct because the modules are represented by 3 LED's, but the topology is, where the voltmeter, power supply, and load is.

This is the strip I am talking about: .
There is a RGB LED in each module, and yes there are 50 modules (LED plus ucs 5903). Right now I can't measure the resistance because it's still on the tree  :) , but if I have aprox 2.7V at the begining and 2.5V at the end, shouldn't the drop be around 0.2V? With the measured 0.7A current it should be around 0.3 ohms (correct me if I am wrong).

About expected current and power...almost no ideea.
The built in resistor should give 18-20 mA (from datasheet) per LED so:
20 * 50 = 1A

The measured current is 0.7A but the voltage is lower than the nominal 5V so I guess it's the correct one.
 

Lurch

  • Guest
Re: Unexpected voltage drop on a low current load
« Reply #8 on: January 02, 2014, 10:56:50 pm »
Just saying nope, won't help me find what is unclear, is clear for me just hard to put it on text (or drawing).

That drawing doesn't make any sense. It is total garbage. Ignoring the fact you don't know how the modules are wired the drawing is a total mess and shows precisely zero amount of clarity or detail.
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #9 on: January 02, 2014, 11:02:03 pm »
Where are you measuring the total current, is it including everything?  Everything is powered from the single 5v supply? 

You haven't shown where the ATMega is powered from.  If you are measuring 2.7 V across the supply at it's output, I don't see how the AT will be powered from that at all.

NB: An RGB led has 3 die in it.  So that could be 3 Amp when driving white.

~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Offline Jon86

  • Frequent Contributor
  • **
  • Posts: 526
  • Country: gb
Re: Unexpected voltage drop on a low current load
« Reply #10 on: January 02, 2014, 11:08:25 pm »
Your drawing's a bit confusing, would you be able to go into more detail with it?
It doesn't really make sense as it is, the ground is at 2.7v and the power is shorted together at two places. Also the data and clock lines go straight to the power?
Death, taxes and diode losses.
 

Offline fyrewurxTopic starter

  • Contributor
  • Posts: 16
Re: Unexpected voltage drop on a low current load
« Reply #11 on: January 02, 2014, 11:27:11 pm »
Just saying nope, won't help me find what is unclear, is clear for me just hard to put it on text (or drawing).

That drawing doesn't make any sense. It is total garbage. Ignoring the fact you don't know how the modules are wired the drawing is a total mess and shows precisely zero amount of clarity or detail.

Lurch, the fact that the strip is doing what I want it to do makes me think that in fact I have a clue about the wiring. I admit that is not a good representation, so from the begining.

Each module has 4 wires, two for serial and two for power. The serial is coming from the MCU (the Atmega8), +5V on the top, on the bottom the ground. The MCU is powered from the same supply as the strip, 5V (in ISIS the power connections for MCU's are not shown). I measured the current before the entire circuit so it's the current for the MCU and modules. The ground is at 0V as it should be and the numbers show what is on the voltmeters. The data and clock don't go to power just very bad drawing on my part. I don't have the library for that chip so I just replaced the entire thing with 3 LED.

The datasheet says that one LED draws 18-20 mA, I assumed that they refer to all dies (about 6 mA per die). And about ATmega8 on 2.7 V, well I would need a scope I think to be sure that it gets that low (cheap dmm can't be thrusted right?).  I'll give it another shot for the drawing, this time tell me what program to use (hope I won't have to learn a new one).
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #12 on: January 02, 2014, 11:38:43 pm »
Your cheap multimeter will be perfectly fine to measure the voltage accurately enough (unless it's battery is nearly flat, or it's been damaged) provided it's fairly constant.

I think that this is effectively your block diagram?  Are you measuring where I have indicated?
~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #13 on: January 02, 2014, 11:40:38 pm »
Also to what datasheet are you referring?
~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Offline fyrewurxTopic starter

  • Contributor
  • Posts: 16
Re: Unexpected voltage drop on a low current load
« Reply #14 on: January 02, 2014, 11:49:02 pm »
Your cheap multimeter will be perfectly fine to measure the voltage accurately enough (unless it's battery is nearly flat, or it's been damaged) provided it's fairly constant.

I think that this is effectively your block diagram?  Are you measuring where I have indicated?

Yes exactly, both current and voltage, I added one more measuring point. I forgot to mention that the LED's are switching color so the current, therefore the voltage is not constant (the startup program is white, red, green, blue at about 0.4 s, not sure if the dmm is fast enough).

The datasheet is this: http://www.leds.ru/pdf/D705_en.pdf, a chip that should be same as the UCS 5903.

Looks like paint is the way to go, i'll use it from now on more.
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #15 on: January 02, 2014, 11:57:36 pm »
I can only put it down to your meter being faulty or your use of it being so (checked the test lead resistance, checked that you have it in the right mode, checked that you have the leads in the right sockets?)

If you are seeing blue OK, well, blue led dies are around 3.2v, I don't think 2.7v would light them.

Your datasheet also specifies a supply voltage of 4.5v minimum for the controllers.

The ATMega won't run on the 2.7v you are measuring.

So, short version, the 2.7v you are measuring at the supply output simply can not be correct if the circuit is otherwise functioning.
~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Lurch

  • Guest
Re: Unexpected voltage drop on a low current load
« Reply #16 on: January 03, 2014, 12:09:52 am »
Looks like paint is the way to go

And it goes downhill from here....
 

Offline fyrewurxTopic starter

  • Contributor
  • Posts: 16
Re: Unexpected voltage drop on a low current load
« Reply #17 on: January 03, 2014, 12:13:27 am »
I can only put it down to your meter being faulty or your use of it being so (checked the test lead resistance, checked that you have it in the right mode, checked that you have the leads in the right sockets?)

If you are seeing blue OK, well, blue led dies are around 3.2v, I don't think 2.7v would light them.

Your datasheet also specifies a supply voltage of 4.5v minimum for the controllers.

The ATMega won't run on the 2.7v you are measuring.

So, short version, the 2.7v you are measuring at the supply output simply can not be correct if the circuit is otherwise functioning.

The meter measures usb at around 5V, the 1.5V batteries at around 1.5V (new ones), 2.5V is 50% error. When I short the probes on ohm, they show around 0.4 ohms, I guess pretty normal. I tryed on multiple ranges (even if I loose digits), the voltage was consistently getting lower.

The circuit is functioning with the PC supply but when I used the 2A wall-plug (from a D-link router) the MCU was definetly reseting (the startup sequence is unique) and the last 10-12 modules were acting up so there is something wrong. Even with PC supply the last LED has different color from the rest of it. I think that without more or different equipment I can't investigate further unless anyone has another ideea.

After checking the datasheet what do you think, the LED's take 20mA or 3*20mA?

Paint might be downhill, but what is the uphill Lurch, what would you use (don't want to repeat this diagram's mistake again)?

Ty for the answers.
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #18 on: January 03, 2014, 12:45:19 am »
=
After checking the datasheet what do you think, the LED's take 20mA or 3*20mA?

My skim reading of the datasheet I think it is 20ma per die (so 60mA total) in it's default configuration without external resistors.

Therefore if all 3 dies are being driven 100%, that's 60mA per module, if you have 50 of these modules on a string, that's 3.6A for the leds themselves.



~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Offline papo

  • Regular Contributor
  • *
  • Posts: 93
  • Country: ch
Re: Unexpected voltage drop on a low current load
« Reply #19 on: January 03, 2014, 04:03:53 am »
It is indeed weird that the circuit runs at 2.7 V for the reasons already stated. I'd like to draw your attention to something else that caught me many times, though: Do not underestimate voltage drops caused by even low resistances. Let's assume the strip is drawing 700 mA. If the (long?) cable connecting the supply to the micro and the strip has a resistance of only 2 Ohms, that's already a 1.4 V drop. To make things worse, you have to double this drop since it occurs not only on the wire supplying the 5 V, but also on the ground return wire. That's 2.8 V of drop right there, of coursed based on the assumption of 2 Ohms per wire for demonstration purposes.

I might add that measuring low resistances with a low end meter might be tricky. But do not underestimate what small resistances can do. For example, I saw someone using two alligator leads to connect a 2 A-DUT to a 5 V supply. These leads had a resistance of 1 Ohm each and it took him a long time to figure out why the DUT rebooted as soon as it drew a considerable amount of current :-(.

This might not be the reason for the weird phenomenon you are expecting, but still something important to keep in mind. Your paint sketch shows 2.5 V at the end of the LED strip. I am not sure whether the 2.7 V are measured at the beginning of the LED strip or at the bare output of the supply. If the former is the case, it might be worth checking the voltage at the supply itself (use both probes at the supply, not e.g. one probe at the 5 V supply output and the other at the ground of the strip). If that sits at 2.7 V as well, either there is something wrong with your measurement or the strip is loading down the supply.

Regards
Matt
 

Offline Huluvu

  • Regular Contributor
  • *
  • Posts: 91
  • Country: de
    • ECM Home
Re: Unexpected voltage drop on a low current load
« Reply #20 on: January 03, 2014, 08:29:59 am »
I would hook up an Oscilloscope to check the voltage (maybe you are reading just an average)


"Yeah, but no, but yeah, but no..."
 

Offline fyrewurxTopic starter

  • Contributor
  • Posts: 16
Re: Unexpected voltage drop on a low current load
« Reply #21 on: January 03, 2014, 10:21:43 am »
Remeasured voltages: at beginning of the strip 2.7 - 2.8V lowest, at the other molex (one molex for output, other for measurement) it's about 2.9-3V.

I guess it's a combination of resistances and unloading of the supply (if the filter caps are old/dry the supply is easier to unload?).

I will definitely measure the resistance of the entire strip (about 3 m) using a constant current and measure the voltage drop. In a real world application if the drop is let's say 1V (+5V wire + ground wire) per 50 modules, you can't drive more then well half of the strip (datasheet says they require 4.5V) right? If your application would require 2, 3 strips what would you do (different strip)?

I would gladly use an oscilloscope but I don't have one. I know that there are bargains on ebay but the shipping here (Romania) is few times the price of the scope.
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #22 on: January 03, 2014, 11:39:03 am »
If your application would require 2, 3 strips what would you do (different strip)?

Inject power at points further down the bus (distribute 12v, step it down to 5v at points of injection).
~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3020
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: Unexpected voltage drop on a low current load
« Reply #23 on: January 03, 2014, 11:46:01 am »
at the other molex

If you're measuring 2.7-2.8v right out of the molex on your power supply under load, your power supply is shagged probably.

You could get yourself some power resistors from a local electronics shop if there is one and do some tests to see how much load you can put on it before it starts going bad, if you can't get 4 amps, you'll have your culprit.
~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Offline fyrewurxTopic starter

  • Contributor
  • Posts: 16
Re: Unexpected voltage drop on a low current load
« Reply #24 on: January 03, 2014, 12:07:37 pm »
sleemanj, thank you for your answers, I will do that (get some resistors - a couple 1 ohm 5W should do it i think). I learned from this thread that when I deal with voltage drops I should check the supply and the wires because they can cause this sort of problems.

One last question: what is your opinion about using ethernet cable for powering circuits, using it on breadboard...well general electronic use?
 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4694
  • Country: au
  • Question Everything... Except This Statement
Re: Unexpected voltage drop on a low current load
« Reply #25 on: January 03, 2014, 01:44:02 pm »
Power over Ethernet is a thing...

but using generic patch cables i would say dont exceed 750mA, and for solid core dont exceed 1A for a 0.5m length

I have a 15m roll of solid core cable that i have been using for breadboarding, though i do have to say the flexible wires with the machined pin crimp style is better long term as the solid core ends will work harden and if your moving stuff often you can have a wire snap off and either cause a hard to track fault or touch something it shouldn't.

Though i only have 1 of those, and for me its difficult to get that style locally :/
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf